Oh shit! SINAD can suck my ****s

Discussion in 'Measurement Techniques Discussion' started by Marvey, Jul 2, 2021.

  1. Cooper32

    Cooper32 Facebook Friend

    Contributor
    Joined:
    Nov 20, 2020
    Likes Received:
    146
    Trophy Points:
    33
    Location:
    Wisconsin, USA
    Another possible set of measurements would be for an OTL amp with both a HD 650 and a low impedance planar. Would be interested to see just how much THD there's at the 3 frequencies for the ideal use case (OTL with high impedance dynamic headphones) vs a less-than-ideal match (OTL with current hungry planars).
     
  2. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,653
    Trophy Points:
    113
    Location:
    On planet
    20210702 HD650 impedance DATS L.png

    perhaps the impedance curve for HD650 might provide enlightenment ...

    possibly 100 Hz will evoke larger differences ...

    also verified on another of my analyzers:

    20210701 HD650 2V FR Dist Imp L.png
     
    • Like Like x 17
    • Epic Epic x 5
    • List
  3. Baten

    Baten Friend

    Pyrate
    Joined:
    Mar 18, 2018
    Likes Received:
    1,130
    Trophy Points:
    93
    Location:
    EU
    Don't believe L30 is using OPA16xx though.
     
  4. Cspirou

    Cspirou They call me Sparky

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    8,200
    Trophy Points:
    113
    Location:
    Northwest France
    Definitely different in terms of back EMF as transistors and transformers are not the same at all. It's why I've always wanted to see amp measurements with real speaker loads instead of transistors.

    I often felt that tube amps (or an ECP amp) often did a lot more than simply add THD but it was never something we measured for.
     
  5. JohnM

    JohnM Author of REW - Rando

    Joined:
    Dec 13, 2015
    Likes Received:
    134
    Trophy Points:
    33
    It may be worth looking at the distortion the headphone is producing with the different amplifiers and whether it is at all different. The output impedance of the amplifier under the test condition should be the main factor determining how much of the load's distortion you see at the amplifier output. The test may not tell you much about the overall effect of amplifier + headphone, however, or the quality of the amplifier.
     
  6. batriq

    batriq Probably has made you smarter

    Pyrate
    Joined:
    Oct 31, 2015
    Likes Received:
    1,041
    Trophy Points:
    93
    I'm not sure if this is possible, but can the table include combinations of two signals? For example:
    • Measure 37Hz + 1kHz 2-frequency signal with resistor load
    • Measure 37Hz + 1kHz 2-frequency signal with headphone load
    • Similarly for 37Hz + 9.3kHz and 1kHz + 9.3kHz
    Hypothesis is that even though the 1kHz SINAD measurement with headphone load is good, the 37Hz + 1kHz measurement will not be. In other words, good SINAD measurements at a single high frequency does not matter if the lower frequency measurement is bad. This is also more realistic since music never comes in a single frequency.
     
  7. 7seven

    7seven Acquaintance

    Joined:
    Oct 28, 2018
    Likes Received:
    56
    Trophy Points:
    28
    Location:
    Ireland
    • Like Like x 3
    • Epic Epic x 1
    • List
    Last edited: Jul 3, 2021
  8. Beefy

    Beefy Friend

    Pyrate
    Joined:
    Apr 10, 2021
    Likes Received:
    1,738
    Trophy Points:
    93
    Location:
    Canada
    Don't read the thread Beefy, it will just make you mad....

    But in all seriousness, and in defence to them, I can understand how it is a lot more difficult to establish a load for speakers, as seems to be discussed in that thread. There is soooooo much more variability in speakers, you have to worry about the interaction with crossovers, and you can expect power outputs over several orders of magnitude. Picking a single reference speaker that 'everyone' has heard would be problematic.

    Headphones are a lot easier to standardise, in that there is a single voice coil, and the range of power outputs isn't nearly as large. But most importantly, settling on a single standardised product like the HD650 that the majority of enthusiasts are familiar with is entirely reasonable. A real load, under real world conditions, no need for simulation.
     
  9. 7seven

    7seven Acquaintance

    Joined:
    Oct 28, 2018
    Likes Received:
    56
    Trophy Points:
    28
    Location:
    Ireland
    That is the very issue I have with it, there is no standardised speaker load so the best option is to not bother and use a basic resistive load? That tells the end user even less than testing a single random speaker alongside the resistive load,
    It certainly would make it more difficult for these manufacturers to design for best possible measurements if they didnt already know the test conditions (a resistor).
    It would also make condensing the measurements down into a single number more difficult, but obviously that is point to sell to the naive, lazy and inexperienced.


    Headphones are easier to standardise, they still show zero interest in them for testing.
     
    Last edited: Jul 3, 2021
  10. Biodegraded

    Biodegraded Friend

    Pyrate Contributor
    Joined:
    May 28, 2017
    Likes Received:
    7,985
    Trophy Points:
    113
    Location:
    Vancouver BC
    But guess what will happen if it turns out that a particular type of headphone (high impedance dynamic) interacts with ASR's favourite type of amp (global feedback, vanishing SINAD) to make the amp behave in a less than favourable way: that entire class of headphone will be declared "badly designed" and will be repeatedly trashed by dear leader and the minions, to the confusion of the rest of the internet.
     
  11. Beefy

    Beefy Friend

    Pyrate
    Joined:
    Apr 10, 2021
    Likes Received:
    1,738
    Trophy Points:
    93
    Location:
    Canada
  12. Cspirou

    Cspirou They call me Sparky

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    8,200
    Trophy Points:
    113
    Location:
    Northwest France
    I would like to bring attention to this blog post by Joseph Crowe (who also posts on ASR). He found that the amp with the lowest IMD with speakers was actually a tube amp.

    https://josephcrowe.com/blogs/news/...-choice-on-speaker-intermodulation-distortion
     
  13. crenca

    crenca Friend

    Pyrate
    Joined:
    May 26, 2017
    Likes Received:
    3,829
    Trophy Points:
    113
    Location:
    Southern New Mexico
    Thinking out loud of counter argument(s) to using the HD650:

    1) Does it not have a lightweight magnet design, relatively speaking, to most other dynamics?
    2) Despite its well deserved position in the hobby, don't most of us "graduate" on to ZMF's, Focal's, various planers, and sometimes small shop/DIY HP's

    Speaking for myself I don't rate the HD650 as much as most here do, but I do admit the Black Amp prototype that is on tour here really 'woke it up' and I heard more linearity and detail out of it than I thought possible (not that it broke out of its fundamental limitations):

    https://austinaudioworks.com/amp-description/

    That amp is a very clean, no feedback, and high current SS design
     
  14. Beefy

    Beefy Friend

    Pyrate
    Joined:
    Apr 10, 2021
    Likes Received:
    1,738
    Trophy Points:
    93
    Location:
    Canada
    Fair warning, useless car analogy incoming.

    Reading reviews of the new Ferrari might be fun. But the reviews that the vast majority of people actually find useful are for the new Toyota Corolla or Mazda 3.
     
  15. crenca

    crenca Friend

    Pyrate
    Joined:
    May 26, 2017
    Likes Received:
    3,829
    Trophy Points:
    113
    Location:
    Southern New Mexico
    On the other hand, is SBAF a sub 3, 4, $500 (pick a number) Topping/Schiit/iFi/JDS/insert_preference_here only kind of place? If the reference is going to be small/cheap/practical amps and HP's only, how are folks going to advance in the hobby? That to my mind is limited - a kind of 'ASR done right' vision that itself is limiting.

    The 'no man's land' of $500-2K is increasingly filled up with good equipment, particularly on the used market. Besides the HD650 just is not what it used to be relatively speaking. To use a car analogy, most of us are purchasing and driving SUV's and trucks now - the Toyota Camry (or any other practical mid-sized sedan) is not as important as it was...
     
  16. Beefy

    Beefy Friend

    Pyrate
    Joined:
    Apr 10, 2021
    Likes Received:
    1,738
    Trophy Points:
    93
    Location:
    Canada
    Yet, are any of us going to pick equipment based on Marv's measurements? No, I really don't think so. We're still going to organise an audition, get on a loaner tour, borrow from a buddy, and buy what sounds best.

    So from an academic interest perspective, I think that reaching the widest possible audience with a gold standard is important.
     
  17. Nimrodor

    Nimrodor New

    Joined:
    Dec 26, 2017
    Likes Received:
    11
    Trophy Points:
    3
    Location:
    California
    The output impedance might have something to do with it, but the diyaudio post isn't applicable here aside from the advice that op-amps are in general not good with capacitive loads. The L30 uses the OPA1612 for the input, which is not a CMOS op-amp and also does not share the same output stage design as the OPA1688/1656/2156. (and I think the open loop output impedance should be dominated by the TPA6120).

    OPA1612+TPA6120. Here's (roughly) how they're connected.

    100Hz might also answer another question about what order the effect is happening at since the maximum magnitude is a zero-crossing for the phase.

    An objective standard should allow results to be easily replicated and compared. For the purpose of measurements, this encourages adoption and keeps everyone honest. The best equipment to standardize on is the one the most people have access to without reducing the quality of the measurement.

    I appreciate this site's use of the HD650 as a subjective standard because it ensures every reviewer is speaking the same language, and that virtually anyone can understand what they mean. If nothing else, the HD650 has proven that it has staying power; older impressions referencing Hifiman, Audeze, MrSpeakers, Abyss, etc. products have since lost value as those companies have discontinued their offerings. The high end tends to be fickle and FOTM-driven, and perhaps that's to be expected; when products are billed as pushing the state of the art, it's understandable that they will be regularly displaced. In some sense those products which advance the hobby may inherently be poor choices to standardize around.
     
  18. Pancakes

    Pancakes Friend

    Pyrate Contributor
    Joined:
    Aug 13, 2020
    Likes Received:
    1,418
    Trophy Points:
    93
    Location:
    Atl
    An interesting discussion about opamps and their variability. I was swapping out op amps on WHAMMY and one of them sounded downright tubey with HD600. Yay. With Elegia, noise floor went through the roof. Not yay.

    I think it's a good idea to include measurements with a handful of cans. Several (3?) dynamic drivers of various price points and a couple planars of different price points.

    I think the HD6 line is a must simply due to its familiarity within the hobby. And then a Clear or Auteur? And maybe something else?

    For planars, the Sundara is a common entry point and is relatively well known. And some other one higher up the tier list.

    I realize this creates a ton more work and requires an ever present collection of cans but I feel like this (Marv's AP adventures and enthusiasm) is a golden opportunity to do things the right way. The comprehensive, fact finding, illuminating way.

    As such, I'll double my Patreon donation when I get to my PC.
     
  19. Beefy

    Beefy Friend

    Pyrate
    Joined:
    Apr 10, 2021
    Likes Received:
    1,738
    Trophy Points:
    93
    Location:
    Canada
    O/T, but you should be really careful rolling OPAMPs in the WHAMMY, particularly BJT input ones, unless you have a scope. Depending on the gain and other factors, you will almost certainly have to optimize the feedback compensation capacitors to make sure it isn't oscillating.

    I'd bet a distorted tube sound and high noise is due to the amp oscillating, and this is really bad for your phones.
     
  20. Grattle

    Grattle Friend

    Pyrate
    Joined:
    Feb 17, 2017
    Likes Received:
    1,003
    Trophy Points:
    93
    Location:
    USA
    Can the equipment be configured to measure IMD of a range, say, 20Hz to 10kHz using a headphone driver or must it be done manually across a selected set of frequencies?
     

Share This Page